-
Bug
-
Resolution: Done
-
Undefined
-
rhel-10.1
-
No
-
Low
-
TestOnly
-
1
-
rhel-pt-c-libs
-
ssg_platform_tools
-
1
-
False
-
False
-
-
No
-
PT Libraries 2025 S09
-
None
-
None
-
Unspecified Release Note Type - Unknown
-
Unspecified
-
Unspecified
-
Unspecified
-
-
ppc64le
-
None
libffi.closure/huge_struct.c testcase segfaults when being run in POWER10 systems. This does not reproduce on POWER9 or older systems.
I was able to reproduce it out of the testsuite and retrieve a backtrace.
To reproduce:
rpmbuild -rc --nocheck libffi-3.4.4-10.el10.src.rpm
cd /root/rpmbuild/BUILD/libffi-3.4.4/powerpc64le-unknown-linux-gnu/testsuite
gcc -g ../../testsuite/libffi.closures/huge_struct.c -W -Wall -Wno-psabi -O2 -mlong-double-128 -I/root/rpmbuild/BUILD/libffi-3.4.4/powerpc64le-unknown-linux-gnu/testsuite/../include -I../../testsuite/../include -I../include/.. -L../.libs -lffi -lm -o ./huge_struct.exe
./huge_struct.exe
Segmentation fault (core dumped)
Backtrace from gdb:
(gdb) run Starting program: /root/rpmbuild/BUILD/libffi-3.4.4/powerpc64le-redhat-linux-gnu/testsuite/huge_struct.exe [Thread debugging using libthread_db enabled] Using host libthread_db library "/lib64/libthread_db.so.1". Program received signal SIGSEGV, Segmentation fault. 0x00007ffff7f27504 in ffi_prep_args64 (ecif=0x7fffffffdd10, stack=<optimized out>) at ../src/powerpc/ffi_linux64.c:768 768 gprvalue = **p_argv.us; Missing separate debuginfos, use: dnf debuginfo-install glibc-2.39-37.el10.ppc64le (gdb) bt #0 0x00007ffff7f27504 in ffi_prep_args64 (ecif=0x7fffffffdd10, stack=<optimized out>) at ../src/powerpc/ffi_linux64.c:768 #1 0x00007ffff7f27ec8 in ffi_call_LINUX64 () at ../src/powerpc/linux64.S:99 #2 0x00007ffff7f23f2c in ffi_call_int (cif=cif@entry=0x7fffffffe600, fn=<optimized out>, fn@entry=0x10001530 <test_large_fn>, rvalue=<optimized out>, rvalue@entry=0x7fffffffe4c0, avalue=avalue@entry=0x7fffffffe180, closure=<optimized out>, closure@entry=0x0) at ../src/powerpc/ffi.c:105 #3 0x00007ffff7f26d00 in ffi_call (cif=cif@entry=0x7fffffffe600, fn=fn@entry=0x10001530 <test_large_fn>, rvalue=rvalue@entry=0x7fffffffe4c0, avalue=avalue@entry=0x7fffffffe180) at ../src/powerpc/ffi.c:138 #4 0x0000000010000ec8 in main (argc=<optimized out>, argv=<optimized out>) at ../../testsuite/libffi.closures/huge_struct.c:298
This was found while testing libffi-3.4.4-10.el10 but it can be reproduced in older libffi-3.4.4-8.el10 as well.
If the test is built without -O2, the segfault does not occur:
# gcc -g ../../testsuite/libffi.closures/huge_struct.c -W -Wall -Wno-psabi -mlong-double-128 -I/root/rpmbuild/BUILD/libffi-3.4.4/powerpc64le-unknown-linux-gnu/testsuite/../include -I../../testsuite/../include -I../include/.. -L../.libs -lffi -lm -o ./huge_struct.exe
# ./huge_struct.exe
1 2 3 4 5 6 7 8 9 10 11 0x12345678 1 2 3 4 5 6 7 8 9 10 11 0x12345678 1 2 3 4 5 6 7 8 9 10 11 0x12345678 1 2 3 4 5 6 7 8 9 10 11 0x12345678 1 2: 2 3 4 5 6 7 8 9 10 11 12 0x12345679 3 4 5 6 7 8 9 10 11 12 13 0x1234567a 4 5 6 7 8 9 10 11 12 13 14 0x1234567b 5 6 7 8 9 10 11 12 13 14 15 0x1234567c 6 7
res: 2 3 4 5 6 7 8 9 10 11 12 0x12345679 3 4 5 6 7 8 9 10 11 12 13 0x1234567a 4 5 6 7 8 9 10 11 12 13 14 0x1234567b 5 6 7 8 9 10 11 12 13 14 15 0x1234567c 6 7
1 2 3 4 5 6 7 8 9 10 11 0x12345678 1 2 3 4 5 6 7 8 9 10 11 0x12345678 1 2 3 4 5 6 7 8 9 10 11 0x12345678 1 2 3 4 5 6 7 8 9 10 11 0x12345678 1 2: 2 3 4 5 6 7 8 9 10 11 12 0x12345679 3 4 5 6 7 8 9 10 11 12 13 0x1234567a 4 5 6 7 8 9 10 11 12 13 14 0x1234567b 5 6 7 8 9 10 11 12 13 14 15 0x1234567c 6 7
res: 2 3 4 5 6 7 8 9 10 11 12 0x12345679 3 4 5 6 7 8 9 10 11 12 13 0x1234567a 4 5 6 7 8 9 10 11 12 13 14 0x1234567b 5 6 7 8 9 10 11 12 13 14 15 0x1234567c 6 7
